Deletion of murine astrocytic vesicular nucleotide transporter increases anxiety and depressive-like behavior and attenuates motivation for reward

Abstract

Astrocytes are multi-functional glial cells in the central nervous system that play critical roles in modulation of metabolism, extracellular ion and neurotransmitter levels, and synaptic plasticity. Astrocyte-derived signaling molecules mediate many of these modulatory functions of astrocytes, including vesicular release of ATP. In the present study, we used a unique genetic mouse model to investigate the functional significance of astrocytic exocytosis of ATP. Using primary cultured astrocytes, we show that loss of vesicular nucleotide transporter (Vnut), a primary transporter responsible for loading cytosolic ATP into the secretory vesicles, dramatically reduces ATP loading into secretory lysosomes and ATP release, without any change in the molecular machinery of exocytosis or total intracellular ATP content. Deletion of astrocytic Vnut in adult mice leads to increased anxiety, depressive-like behaviors, and decreased motivation for reward, especially in females, without significant impact on food intake, systemic glucose metabolism, cognition, or sociability. These behavioral alterations are associated with significant decreases in the basal extracellular dopamine levels in the nucleus accumbens. Likewise, ex vivo brain slices from these mice show a strong trend toward a reduction in evoked dopamine release in the nucleus accumbens. Mechanistically, the reduced dopamine signaling we observed is likely due to an increased expression of monoamine oxidases. Together, these data demonstrate a key modulatory role of astrocytic exocytosis of ATP in anxiety, depressive-like behavior, and motivation for reward, by regulating the mesolimbic dopamine circuitry.

Introduction

The brain is a highly heterogeneous organ in the body, containing many neuronal and non-neuronal cell types with highly specialized cellular functions. Among them, astrocytes are the most abundant glial cell in the brain. Each astrocyte has hundreds of fine processes, which enwrap neuronal cell bodies, dendrites and synapses, as well as brain vasculatures^{1, 2}. These processes allow each astrocyte to sense and respond to minor fluctuations of environmental cues within its anatomically unique and spatially non-overlapping domain³. Thus, astrocytes show a high degree of metabolic flexibility to adapt to both systemic energy states and neural activities^4–6. Astrocytes rapidly take up excess glutamate and GABA through astrocyte-specific glutamate transporters (GLAST and GLT-1) and GABA transporter (GAT)^7–11, to ensure the temporal and spatial regulation of synaptic activity. In addition, astrocytes become reactive in many pathophysiological conditions like neurodegeneration, trauma, and ischemia^12–14. The functional alterations in reactive astrocytes under these pathophysiological conditions may further change the course of disease progression¹⁵.

A growing body of evidence has demonstrated that astrocytes are an active modulator within the neural circuitry. Astrocytes can release metabolites, small peptides, and neurotransmitters like glutamate, D-serine, and ATP¹⁶. This “gliotransmission” enables astrocytes to communicate with nearby neurons to regulate synaptic activity and plasticity^17–21. Astrocyte-derived ATP molecules exert their effects on nearby neurons and glial cells via activation of a collection of ionotropic (P2X) and metabotropic (P2Y) purinergic receptors. In addition, extracellular ATP is quickly hydrolyzed and converted to adenosine, which can also signal through selective adenosine receptors. Collectively, astrocyte-derived ATP and related purinergic molecules have been shown to play important roles in multiple neural circuits. Thus, our group has shown that insulin induces exocytosis of ATP in astrocytes to potentiate dopamine release in the nucleus accumbens²². Others have shown that purinergic signaling molecules derived from astrocytes modulate various neural circuits and relevant behaviors, including cognition, mood, appetite, sleep, and somatosensation^{19, 20, 23–27}.

The release of ATP by astrocytes may involve selective channels, non-selective gap junction hemichannels, and regulated exocytosis^28–30. The relative contribution and importance of these ATP releasing routes in astrocytes, however, is not completely understood. Slc17a9, also called vesicular nucleotide transporter (Vnut), is an inorganic phosphate transporter responsible for loading cytosolic ATP into secretory vesicles in the presence of Cl^{− 31}. Vnut is widely expressed in neurons, astrocytes, and microglia in the central nervous system, as well as many other peripheral tissues, indicating a general mechanism of cellular ATP release through exocytosis of secretory vesicles. Supporting the functional significance of this cellular machinery, studies targeting Vnut gene have demonstrated a faciliatory role of insulin secretion in pancreatic β-cells during both basal and glucose-stimulated states³². In addition, Vnut-dependent ATP release from dorsal horn neurons has been shown to contribute to neuropathic and inflammatory pain^{33, 34}. The role of Vnut in astrocytes, in contrast, remains largely uncharacterized.

In the present study, we utilized the lox-cre recombination approach to target the mouse Vnut gene with temporal and cell-type specificity. We show that loss of Vnut reduces ATP release by ~50% in primary cultured astrocytes without affecting the general molecular machinery for exocytosis. Astrocyte-specific Vnut knockout mice display normal systemic metabolism, as demonstrated by unaltered body weight, serum glucose and insulin levels, glucose tolerance and insulin tolerance compared with control littermates, as well as no apparent impact on cognitive functions. On the other hand, loss of Vnut in astrocytes results in significantly increased anxiety and depressive-like behavior, especially in female mice, and this is associated with a reduction in motivation for reward and, at a mechanistic level, reduced basal extracellular dopamine levels in the nucleus accumbens. Together, these data demonstrate a key role of astrocytic exocytosis of ATP in modulating dopaminergic neural circuit and reward-related behavior.

Materials and Methods

Ethics approval.

All animal studies were conducted in compliance with the regulations and ethics guidelines of the NIH and were approved by the IACUCs at New York Institute of Technology College of Osteopathic Medicine (2023-WKC-01), Joslin Diabetes Center (97–05), and Tufts Medical Center / Tufts University School of Medicine (B2021–125).

Animals.

Mice were housed in standard cages with a 12-hour light/12-hour dark cycle and fed a chow diet (Rodent Diet 500I, LabDiet). All animal studies were conducted in compliance with the regulations and ethics guidelines of the NIH and were approved by the IACUCs at New York Institute of Technology College of Osteopathic Medicine, Joslin Diabetes Center, and Tufts Medical Center / Tufts University School of Medicine. Vnut^f/f mice were generated by Dr. Bradford Lowell’s lab at BIDMC, Harvard Medical School. Briefly, the Vnut (gene symbol, Slc17a9) gene targeting construct was prepared using recombineering technology³⁵ and a mouse 129 BAC genomic clone bearing the Vnut gene. The upstream and downstream LoxP sites were inserted, respectively, into introns one and two of the Vnut gene. This targeting construct, which also contained a Frt-flanked neomycin cassette for positive selection, was then electroporated into ES cells (W4/129S6, Taconic Biosciences). Correctly targeted neomycin-resistant ES clones were then identified and injected into C57BL6 blastocysts. The resulting chimeric offspring where then bred for germline transmission of the targeted allele. Finally, the Frt-flanked neomycin resistance cassette was deleted by crossing with mice with germline expression of Flp recombinase. The Vnut^f/f mice were then backcrossed to a C57BL/6J background for 10 generations. To delete Vnut in astrocytes in adult animals, Vnut^f/f mice were crossed with GFAP-CreERT2 mice (JAX# 012849) or Aldh1l1-CreERT2 mice (JAX# 031008). For the induction of CreERT2-mediated flox allele recombination, mice were daily injected (i.p.) with 100 mg/kg tamoxifen (Sigma) dissolved in 10% ethanol and 90% peanut oil (Sigma) for 5 days at 6 weeks of age. Vnut^f/f littermates received the same tamoxifen injections to serve as controls. Data from both male and female mice were included in analyses. The sex and the number of animals for each experiment are indicated in the results and figure legends.

Reagents and materials.

Adenovirus encoding CMV promoter-driven GFP:Cre, GFP alone, Cre alone, or Luciferase were purchased from Vector Biolabs and amplified by the Viral Core at NYITCOM. Adenovirus encoding CMV promoter-driven mouse Vnut-mCherry fusion protein was produced by VectorBuilder. AAV8-GFAP-GFP was also produced by VectorBuilder. AAV8-GFAP-GFP:Cre was obtained from UNC Vector Core. Mant-ATP, Lysotracker Red, and Goat anti-Munc18c antibody (#PA5-37892) were purchased from ThermoFrisher Scientific. Rat anti-Lamp1 (ab25245), rabbit anti-S100β (#ab52642), rabbit anti-MaoA (ab126751), and rabbit anti-MaoB (ab259928) were purchased from Abcam. Mouse anti-GFAP (#MAB306) and guinea pig anti-VNUT (#ABN83) was purchased from Millipore. Rabbit anti-c-fos (#2250), rabbit anti-Vamp7 (#14811), and rabbit anti-GAPDH (#5174) were from Cell Signaling Technology. Rabbit anti-syntaxin-4 (#110041) was from Synaptic Systems. Rabbit anti-TH (NB300-109) was from Novus. HRP-conjugated goat-anti-rabbit IgG secondary antibodies were purchased from GE Healthcare. Alexa fluoro-dye conjugated secondary antibodies for immunofluorescence studies were all purchased from ThermoFisher Scientific.

Primary astrocyte cultures.

Primary cortical astrocyte cultures were prepared from Vnut^f/f newborn pups as previously described²². Briefly, cortices were surgically dissected from newborn pups, and pooled for the following digestion in Hibernate-A media supplemented with 4 mg/ml papain (Sigma) and 33 U/ml DNase I (Sigma) on a shaker at 37°C for 30 min. Dissociated cells were plated on a T75 flask and cultured in DMEM/F12 (Gibco) plus 10% FBS and 1x pen/strep (Gibco). Non-astrocytes were depleted by vigorously shaking the following day. Culture media was changed every 3 days. Upon confluency, astrocytes were trypsinized and replated onto 12-well plates (Corning) or 8-chamber slides (ibidi). To induce flox allele recombination, Vnut-flox astrocytes were infected with adenovirus encoding Cre:GFP (1 × 10⁹ GC/ml) overnight, and cultured for an additional 5 days before experiments. To generate the control cells, the same floxed astrocytes were infected with adenovirus encoding GFP alone. For studies involving fluorescent labeling, adenovirus encoding Cre or Luciferase were used to induce Vnut deletion or serve as controls.

ATP release.

Primary astrocytes cultured in 12-well plates were washed twice with 1x HBSS and replaced with 1x HBSS supplemented with 100 μM ARL67156 (an ecto-ATPase inhibitor). After 30 minutes of incubation, the supernatant was collected, and ATP quantitated using a luciferase-based ATP determination kit (Thermo Fisher Scientific). Cells were lysed with RIPA lysis buffer complemented with 1x protease cocktail (Bimake). Protein concentrations were determined using a BCA kit (ThermoFisher Scientific) and used for normalization. To quantify the total ATP content in the cells, 50 μl of cell lysates were mixed with 50 μl 2 M perchloric acid by vortexing and incubated on ice for 5 minutes. After centrifugation, the supernatant was mixed with 50 μl 3 M KOH, vortexed, and centrifuged to precipitate the remaining perchloric acid. The resultant supernatant was neutralized using HCl and subjected to ATP assay as described above.

Tissue section preparation and immunofluorescence.

Mice were anesthetized with an intraperitoneal (i.p.) injection of Ketamine (100 mg/kg)/Xylazine (20 mg/kg) mix, transcardially perfused with ice cold 1X PBS followed by 10% buffered formalin and decapitated. After 24 h of post-fixation in 10% buffered formalin, brains were removed from the skull, post-fixed for an additional 24 h, cryoprotected in 30% sucrose, and quickly frozen on dry ice. Serial coronal 30 μm sections were cut using a Cryostat. Brain sections for immunostaining were placed in 24-well plates and washed three times in 1x PBS for 5 min. The sections were then permeabilized and blocked by the addition of 5% normal goat serum (NGS) and 0.1% Triton X-100 in 1x PBS and agitated for 30 min at RT. After blocking, primary antibodies diluted in 1x PBS containing 1% NGS + 0.1% Triton X-100 were added with gentle rocking overnight at 4°C. On the following day, sections were washed four times with 1x PBST and secondary antibody diluted in 1x PBS containing 1% NGS + 0.1% Triton X-100 was applied. Sections were then washed four times with 1x PBST and coverslipped with SlowFade Gold containing DAPI (Invitrogen). Images were captured by a Zeiss LSM 980 with Airyscan 2 confocal microscope or a Zeiss Axio Observer microscope. The mean intensity and pixel areas of the confocal images were analyzed using ImageJ software.

Quantification of c-fos⁺ neurons in nucleus accumbens.

Female Vnut^f/f and Vnut^Gfap KO mice at basal condition or 1 h post forced swimming test (FST) were sacrificed, and their brains collected, fixed in 10% formalin, cryoprotected and cryosectioning as described above. The brain sections were co-stained for c-fos and NeuN. Tiled confocal images were captured to cover both nucleus accumbens core and shell for each mouse using 20x objective. Total numbers of c-fos⁺ /NeuN⁺ cells and NeuN⁺ cells in the NAc Core and NAc Shell were counted using ImageJ and presented as percentage of c-fos⁺ NeuN⁺ / total NeuN⁺ cells.

Behavioral assessment.

Male and female (4–5 months of age) mice were used for behavioral tests. Each mouse was exposed to each specific behavioral test only once in its lifetime unless otherwise indicated. Only 1 behavioral test was performed per week to minimize the stress.

Open field test:

Each mouse was placed in an open box 57 × 37 × 31 cm. During the 5-minute session, the movement of the mouse was recorded with an HD webcam and analyzed by ANY-maze software (Stoelting Co.). Central zone entries, total distance traveled, and maximum speed were measured.

Forced swimming test:

Each mouse was placed in a vertical Plexiglas cylinder (40 cm height, 18 cm diameter) containing water 15 cm deep (23°C–25°C). The mouse was allowed to swim inside the cylinder and videotaped for 6 minutes. Videos were analyzed by ANY-maze software. During the testing period, the total time of immobility was assessed as a measure of “despair”-like or depressive-like behavior.

Sucrose preference test:

Mice were single-housed and habituated for 24 hours in the behavioral testing room with both bottles containing water. After habituation, the mouse was given free access to 2 water bottles: one containing 1% sucrose solution and the other plain water. Daily water and sucrose intake was measured for the next 4 days. The placement of water and sucrose bottles was switched each day to avoid memory-driven behavior. Sucrose preference was presented as the percentage of the volume of sucrose intake over the volume of total fluid intake.

Effort-based motivational behavior:

Feeding Experimentation Device (FED3, Open Ephys), an open-source device, was used to conduct an operant task and assess reward in mouse home cages³⁶. Prior to training, mice were single housed for 3 days to habituate. 24-h food intake of each mouse was measured. Before each day’s training session, mice were slightly food restricted overnight with 50% of their daily ration. All animals were trained on fixed ratio 1 (FR1) schedule for 10 days, during which mice learned to nose-poke the active portal for a 20 mg sucrose pellet (Bio-Serv). Then, all animals were trained on fixed ratio 5 (FR5) schedule for additional 5 days, in which 5 nose-pokes on the active portal are required for the dispense of one 20 mg sucrose pellet (Bio-Serv). Each training session of FR1 and FR5 lasted 1 hour or till the testing animal consumed 30 pellets. To assess the maximal effort the mice were willing to spend to acquire sucrose pellets, animals were switched to progressive ratio (PR) schedule, during which the response requirement for each reinforcement during the session was calculated according the formula 5e^(reinforcement number × 0.2) − 5, rounded to the nearest integer³⁷. The break point, defined as the last ratio level completed before mice give up in the 90-min session of each animal was recorded. Average break points of 3 sessions were presented as a quantification of motivation for reward.

Evoked dopamine release measured by carbon fiber amperometry.

4-month-old female Vnut^f/f and Vnut^Gfap KO mice were euthanized using a Ketamine (100 mg/kg)/Xylazine (20 mg/kg) mix cocktail. The brain was placed in ice-cold oxygenated sucrose bicarbonate solution (210 mM sucrose, 10 mM glucose, 3.5 mM KCl, 1 mM CaCl₂, 4 mM MgCl₂, 1.25 mM NaH₂PO₄). The hemispheres of the neocortex were promptly glued to a metallic base fitting a Leica VT1000S Vibratome (Leica Microsystems) and cut in 300-μm coronal brain slices, which were transferred to a container filled with oxygenated aCSF (124 mM NaCl, 2 mM KCl, 1.25 mM KH₂PO₄, 2 mM MgSO₄, 25 mM NaHCO₃, 2 mM CaCl₂, and 11 mM glucose) at room temperature. Brain slices were allowed to recover for an hour after dissection. Slices with the 3 sites that contain the bulk of terminals projecting from the dopaminergic midbrain (prefrontal cortex, NAc, or CPu) were used for testing.

Amperometric electrodes were 5-μm carbon fibers (Amoco). The electrodes were backfilled with 3 M KCl and beveled at the tip. Electrode response was tested through their real-time response to an external dopamine standard through amperometry and/or cyclic voltammetry. A positive 700-mV voltage (vs. an Ag-AgCl ground) was applied to the carbon fiber electrode using a 200B amplifier (Axon Instruments). The amperometric electrode was placed in the prefrontal cortex, dorsal striatum, or NAc. A bipolar stimulating electrode (PlasticsOne) was placed 100–200 μm away from the carbon fiber electrode. A current stimulus of +500 μA was applied 5 times per site every 5 minutes for 2 milliseconds. The response of the amperometric electrode (increase above baseline) was recorded using Axograph. The output was digitized at 50 kHz, low-pass-filtered at 1 kHz, and analyzed using Axograph. The number of molecules oxidized was determined by the relation N = Q/nF, where Q is the charge of the spike, n is the number of electrons transferred (2 for catecholamines)³⁸, N is the number of moles, and F is Faraday’s constant (96,485 coulombs per unit charge).

Stereotaxic AAV injections.

Female Vnut^f/f mice were anesthetized with a ketamine (100 mg/ml) and xylazine (20 mg/ml) mix diluted in saline and placed into a stereotaxic station (Stoelting). The skull was exposed via a small incision, and a small hole was drilled to the desired coordinates for viral injection. A 33-gauge needle (Hamilton) was inserted bilaterally into the NAc (from bregma: anterior–posterior, +1.45 mm; medial–lateral, ±0.7 mm; dorsal–ventral, − 4.0 mm from dura). AAV8-GFAP-Cre:GFP or GFP-only (0.5 μl, 0.1 μl/min) were injected bilaterally into NAc using a 10-μl Hamilton microsyringe. The needle was withdrawn 5 minutes after the injection. Animals were assessed for behavior 2 weeks after surgeries.

Intracerebroventricular injection.

Both 4-month-old female Vnut^f/f and Vnut^Gfap KO mice were anesthetized with a ketamine (100 mg/ml) and xylazine (20 mg/ml) mix diluted in saline and placed into a stereotaxic station (Stoelting). A stainless steel guide cannula (PlasticsOne) with 2 mm guide length was implanted into the lateral cerebral ventricle. The stereotaxic coordinates were −0.5 mm posterior and +1.0 mm lateral from bregma. Correct placement of the cannula was verified by i.c.v. injection of angiotensin II (1 μg in 1 μl 0.9% saline; Sigma-Aldrich) 3 days after the cannula implantation. Mice that failed to consume water within 30 minutes after the injections were removed from the following study. One week after the implantation, mice were injected with 2 μl saline or 2-Me-SATP (10 μM) i.c.v. in a randomized fashion 1 hour before the forced swimming test. One group of mice were given saline followed by forced swimming test at week 1 and then given 2-Me-SATP followed by forced swimming test at week 2. In the second group of mice the order of tests was reversed.

Microdialysis.

Both 5-month-old female Vnut^f/f and Vnut^Aldh1l1KO mice were anesthetized with a ketamine (100 mg/ml) and xylazine (20 mg/ml) mix diluted in saline and placed into a stereotaxic station (Stoelting). A guide cannula (CSG-4, 4 mm shaft length, Amuza) was implanted at anterior–posterior, +1.5 mm, medial–lateral, −0.7 mm from bregma, and capped by a dummy cannula. Two weeks after the surgery, a CX-I-4–1 microdialysis probe (cut off: MW 50 kDa, active membrane: 1 mm, Amuza) was inserted into the guide cannula. The experimental mouse was placed in a cage with access to food and water ad libitum, and the microdialysis probe was perfused with aCSF at a flow rate of 0.5 μL/min. After 3 h of perfusion, dialysate fractions were collected at 30 min intervals for 2 h. Dialysates and dopamine standards were subjected to dopamine quantification using the HTEC-ECD 510 system (Amuza).

Data availability.

The present study does not include any omics data to be deposited to public repository. All the raw images and raw data will be provided upon request for research purpose only.

Statistical analysis.

All the data are presented as mean ± SEM. Two independent groups were compared using unpaired 2-tailed t-test. Two-way ANOVA was performed to detect the interactions between genotype and treatment, and Tukey’s post hoc analysis was performed when appropriate. A P value less than 0.05 was considered significant.

Results

Deletion of Vnut in primary astrocytes.

Vnut is a multi-transmembrane transporter located in secretory vesicles in cells. In particular, Vnut is localized in secretory lysosomes in astrocytes. Thus, Vnut-mCherry fusion proteins predominantly co-localized with lysosomal marker Lamp1 in primary cultured astrocytes (Fig. 1A). To investigate the cell-autonomous functions of Vnut in astrocytes, we infected primary cultured cortical astrocytes from Vnut^f/f new-born pups with adenovirus encoding Cre:GFP fusion protein. As controls, Vnut^f/f astrocytes were infected with adenovirus encoding GFP alone. The messenger RNA sequence corresponding to the floxed exon 1 of Vnut gene was undetectable in Vnut^f/f astrocytes following Cre:GFP expression, demonstrating efficient Cre-mediated recombination (Fig. 1B). We named these cells VnutKO hereafter.

Figure 1. Vnut deletion in primary astrocytes does not cause major changes in the expression of signature astrocyte marker genes.

(A) Fluorescent images of primary astrocytes expressing mCherry-fused Vnut protein (red) and immunostained with Lamp1 (green). DAPI (blue) labels the nuclei of the astrocytes. Scale bar: 10 μm. (B) Reverse-transcription PCR to examine the expression of the exon 1 of the Vnut mRNA in astrocytes infected with adenovirus encoding GFP or GFP:Cre fusion protein. The expression of Tbp mRNA was used as loading controls. (C) Relative mRNA expression of signature genes for astrocytes in control and VnutKO astrocytes. Data are shown as mean ± SEM. Two-tailed t-test. *, P < 0.05; **, P < 0.01. N = 4. (D) Relative mRNA expression of Lamp1 and Lamp2 in control and VnutKO astrocytes. Data are shown as mean ± SEM. N = 4. (E) Immunoblotting showing the protein expression of Gfap and Lamp1 in control and VnutKO astrocytes. Gapdh is used as loading control. (F) Densitometric analysis of Gfap and Lamp1 in control and VnutKO astrocytes. Data are shown as mean ± SEM. N = 6. (G) Immunostaining of Lamp1 (red) in control and VnutKO astrocytes. DAPI (blue) labels the nuclei of the astrocytes. Scale bar: 10 μm. (H) Mean intensity of Lamp1 immunofluorescent signal of all particles in control and VnutKO astrocytes. Data are shown as mean ± SEM. Two-tailed t-test. **, P < 0.01. N = 5 for Vnut^f/f and 4 for VnutKO coverslips.

Loss of Vnut did not significantly alter the transcription levels of astrocyte signature genes, including Gfap (glial fibrillary acidic protein), Glul (glutamine synthetase), Gja1 (Connexin-43), and Apoe (apolipoprotein E) (Fig. 1C). The mRNA expression of glutamate transporters, Glt-1 and Glast, showed a modest yet significant reduction in VnutKO astrocytes (Fig. 1C). Interestingly, the expression of Aquaporin 4 (Aqp4) was reduced by almost 50% in VnutKO astrocytes (Fig. 1C).

Given the primary localization of Vnut protein in lysosomes in astrocytes³⁹, we next examined the effect of Vnut loss on lysosomes in primary astrocytes. Loss of Vnut did not affect the transcription of lysosomal proteins, Lamp1 and Lamp2 (Fig. 1D). The protein levels of Lamp1 and astrocyte marker protein Gfap were also comparable between control and VnutKO astrocytes (Fig. 1E, F). Interestingly, of all the Lamp1⁺ vesicles, the mean fluorescent intensity of Lamp1 was reduced by 30% in VnutKO astrocytes (Fig. 1G, H). These data indicate that although the total protein amount of Lamp1 is not changed in VnutKO astrocytes, the intracellular localization and distribution of Lamp1 may be altered.

Loss of Vnut impairs ATP release in primary astrocytes.

Vnut is responsible for loading cytosolic ATP into vesicles in a membrane potential- and Cl⁻-dependent manner³¹. To assess the lysosomal loading of ATP in astrocytes, both control and VnutKO astrocytes were incubated with fluorescent ATP analog Mant-ATP. To visualize lysosomes, astrocytes were loaded with Lysotracker Red. As expected, Mant-ATP was primarily enriched in Lysotracker-labeled lysosomes in Vnut^f/f astrocytes (Fig. 2A). In contrast, the signal intensity of Mant-ATP in lysosome was significantly reduced by almost 70% in VnutKO astrocytes (Fig. 2A, B), demonstrating impaired lysosomal ATP loading in the absence of Vnut in astrocytes.

Figure 2. Loss of Vnut impairs ATP loading into secretory lysosomes and ATP release in primary astrocytes.

(A) Representative confocal images showing intracellular localization of ATP analog Mant-ATP (cyan) and Lysotracker Red in control and VnutKO astrocytes. Scale bar: 20 μm. (B) Quantification showing the mean fluorescent intensity of Mant-ATP in Lysotracker Red-labeled vesicles in control and VnutKO astrocytes. Data are shown as mean ± SEM. Two-tailed t-test. ***, P < 0.001. N = 5 coverslips. (C) ATP release from control and VnutKO astrocytes in 1x HBSS for 30 min normalized to total protein content. Data are shown as mean ± SEM. Two-tailed t-test. **, P < 0.01. N = 6. (D) Total intracellular ATP content in control and VnutKO astrocytes normalized to total protein content. Data are shown as mean ± SEM. N = 6. (E) Relative mRNA expression of SNARE proteins enriched in astrocytes in control and VnutKO astrocytes. Data are shown as mean ± SEM. Two-tailed t-test. *, P < 0.05; N = 4. (F) Immunoblotting showing the protein expression of Stx4, Vamp7 and Munc18c in control and VnutKO astrocytes. Gapdh is used as loading control. (G) Densitometric analysis of Stx4, Vamp7 and Munc18c in in control and VnutKO astrocytes. Data are shown as mean ± SEM. Two-tailed t-test. *, P < 0.05; N = 6.

ATP release from VnutKO astrocytes was decreased by ~50% compared with control Vnut^f/f astrocytes infected with adenovirus encoding GFP alone (Fig. 2C), further demonstrating a functional impact of Vnut loss in astrocytes. This reduction was not due to any changes in total intracellular content of ATP in VnutKO astrocytes (Fig. 2D). Notably, there were no changes in the mRNA expression of key astrocyte-enriched SNARE complex proteins responsible for regulated exocytosis, including syntaxin-4, Snap23, Vamp3/7, and Munc18a/b, and synaptotagmin 11 (Fig. 2E). Munc18c showed a small reduction at messenger level, but not at protein level (Fig. 2E–G). Despite a lack of change in mRNA levels, VnutKO astrocytes showed a ~25% reduction in protein level of syntaxin-4 (Fig. 2F, G). Together, loss of Vnut significantly impairs ATP release in primary astrocytes, without a dramatic impact on the expression of the core SNARE complex machinery.

Generation of the astrocyte-specific Vnut knockout mouse model.

To investigate the impact of Vnut loss in astrocytes in vivo, Vnut^f/f mice were crossed with astrocyte-specific GFAP-CreERT2 mice. Both Vnut^f/f and Vnut^GFAPmice received daily tamoxifen injections (100 mg/kg, i.p.) for 5 days at the age of 6 weeks. 6 weeks post tamoxifen injections, astrocytes and microglia were isolated from mouse brains using FACS⁴⁰. Using qPCR primers specific to the exon 1 of Vnut gene, we observed ~45% reduction of Vnut mRNA expression in astrocytes isolated from Vnut^GFAPKO mice (Fig. 3A), whereas the expression of Vnut in isolated microglia from the same mice was not changed (Fig. 3B). Consistently, the majority of the S100β⁺ astrocytes showed no detectable expression of Vnut protein in Vnut^GFAPKO mouse brain sections (Fig. 3C, arrows), while some S100β⁺ astrocytes preserved the expression of Vnut protein (Fig. 3C, arrowheads), in line with the expected recombination efficiency by GFAP-CreERT2 driver.

Figure 3. Loss of Vnut in astrocytes in adult mice shows minor effects on the expression of astrocyte marker genes, as well as genes involved in exocytosis and lysosomes.

(A) Relative mRNA expression of Vnut in astrocytes FACS sorted from 4-month-old female Vnut^f/f and Vnut^GfapKO mice. Data are shown as mean ± SEM. Two-tailed t-test. *, P < 0.05; N = 4. (B) Relative mRNA expression of Vnut in microglia FACS sorted from 4-month-old female Vnut^f/f and Vnut^GfapKO mice. Data are shown as mean ± SEM. N = 4. (C) Representative confocal images of 4-month-old female Vnut^GfapKO mouse brain sections stained with Vnut (red) and astrocyte-specific marker S100β (green). Arrows: astrocytes without detectable Vnut expression; Arrowheads: astrocytes with Vnut expression. Scale bar: 20 μm. (D) Relative mRNA expression of astrocytes signature genes in astrocytes FACS sorted from 4-month-old female Vnut^f/f and Vnut^GfapKO mice. Data are shown as mean ± SEM. Two-tailed t-test. *, P < 0.05; N = 4 for Vnut^f/f and 5 for Vnut^GfapKO. (E) Relative mRNA expression of SNARE proteins in astrocytes FACS sorted from 4-month-old female Vnut^f/f and Vnut^GfapKO mice. Data are shown as mean ± SEM. N = 4 for Vnut^f/f and 5 for Vnut^GfapKO. (F) Relative mRNA expression of Lamp1 and Lamp2 in astrocytes MACS sorted from 4-month-old female Vnut^f/f and Vnut^GfapKO mice. Data are shown as mean ± SEM. N = 6. (G) Representative confocal images of nucleus accumbens core of the 4-month-old female Vnut^f/f and Vnut^GfapKO mouse brain stained with Lamp1 (red) and astrocyte-specific marker S100β (green). Scale bar: 10 μm. (H) Quantification showing the percentage of Lamp1⁺ particles in S100β⁺ astrocytes. Data are shown as mean ± SEM. Two-tailed t-test. *, P < 0.05. N = 44 cells from 4 animals.

Vnut^GFAP KO mouse showed no apparent malformation in the brain (Supplemental Fig. 1). Loss of Vnut in astrocytes in adult mice had little impact on the expression of astrocyte signature genes, except for Glast which was significantly but modestly upregulated by 30% (Fig. 3D). In addition, the expression of all the major SNARE complex proteins in the isolated astrocytes from Vnut^GFAPKO mouse brains were comparable with those from Vnut^f/f mouse brains (Fig. 3E). Intriguingly, while the mRNA expression of Lamp1 and Lamp2 was not altered in Vnut^GFAPKO astrocytes (Fig. 3F), the Lamp1⁺ area in S100β⁺ astrocytes was reduced by 15% in Vnut^GFAPKO mouse brains compared with that in Vnut^f/f mouse brains (Fig. 3G, H). This finding again suggests altered intracellular localization and distribution of Lamp1⁺ lysosomes in astrocytes that lack Vnut expression.

Astrocyte-specific Vnut knockout on systemic glucose metabolism and insulin sensitivity.

Astrocytes have been shown to regulate systemic metabolism^{41, 42}. We next sought to examine the systemic glucose homeostasis and insulin sensitivity in Vnut^GFAPKO mice. Thus, 4-month-old male Vnut^GFAPKO mice displayed similar body weight, fasting blood glucose, and fasting serum insulin levels compared with Vnut^f/f littermates (Fig. 4A–C). In response to i.p. glucose injections, Vnut^GFAPKO mice displayed an acute elevation of blood glucose levels which plateaued during 15–30min, and a gradual decline to baseline similar to Vnut^f/f littermates (Fig. 4D). In addition, Vnut^GFAPKO mice showed comparable glucose lowering effect as Vnut^f/f littermates in response to exogenous insulin administration (1 mU/g, i.p.) (Fig. 4E). Similarly, 4-month-old female Vnut^GFAPKO mice showed similar body weight, fasting glucose and insulin levels, glucose tolerance, and insulin sensitivity as Vnut^f/f littermates (Fig. 4F–J). Together, loss of Vnut in astrocytes in adult mice does not impact systemic glucose homeostasis and insulin sensitivity.

Figure 4. Loss of Vnut in astrocytes in adult mice does not affect systemic glucose metabolism and insulin sensitivity.

(A) The body weight of 4-month-old male Vnut^f/f and Vnut^GfapKO mice. Data are shown as mean ± SEM. N = 7 for Vnut^f/f and 8 for Vnut^GfapKO. (B) Serum glucose levels of 4-month-old male Vnut^f/f and Vnut^GfapKO mice after overnight fasting. Data are shown as mean ± SEM. N = 7 for Vnut^f/f and 8 for Vnut^GfapKO. (C) Serum insulin levels of 4-month-old male Vnut^f/f and Vnut^GfapKO mice after overnight fasting. Data are shown as mean ± SEM. N = 7 for Vnut^f/f and 8 for Vnut^GfapKO. (D) Glucose tolerance test of 4-month-old male Vnut^f/f and Vnut^GfapKO mice after overnight fasting. Data are shown as mean ± SEM. N = 7 for Vnut^f/f and 8 for Vnut^GfapKO. (E) Insulin tolerance test of 4-month-old male Vnut^f/f and Vnut^GfapKO mice after 4 h fasting early in the morning. Data are shown as mean ± SEM. N = 6 for Vnut^f/f and 7 for Vnut^GfapKO. (F) The body weight of 4-month-old female Vnut^f/f and Vnut^GfapKO mice. Data are shown as mean ± SEM. N = 8 for Vnut^f/f and 7 for Vnut^GfapKO. (G) Serum glucose levels of 4-month-old female Vnut^f/f and Vnut^GfapKO mice after overnight fasting. Data are shown as mean ± SEM. N = 8 for Vnut^f/f and 7 for Vnut^GfapKO. (H) Serum insulin levels of 4-month-old female Vnut^f/f and Vnut^GfapKO mice after overnight fasting. Data are shown as mean ± SEM. N = 8 for Vnut^f/f and 7 for Vnut^GfapKO. (I) Glucose tolerance test of 4-month-old female Vnut^f/f and Vnut^GfapKO mice after overnight fasting. Data are shown as mean ± SEM. N = N = 8 for Vnut^f/f and 7 for Vnut^GfapKO. (J) Insulin tolerance test of 4-month-old female Vnut^f/f and Vnut^GfapKO mice after 4 h fasting early in the morning. Data are shown as mean ± SEM. N = 8 for Vnut^f/f and 7 for Vnut^GfapKO.

Astrocyte-specific Vnut knockout on anxiety and depressive-like behavior.

Astrocyte-derived ATP has been shown to elicit anxiolytic and anti-depressive effect in mice²⁵. With our unique mouse model deficient in astrocytic exocytosis of ATP, we next sought to determine the functional significance of this pathway on behavior. Thus, we assessed the anxiety and depressive-like behaviors in Vnut^GFAPKO mice. During an open field test, 4-month-old female mice displayed almost 50% reduction in center zone entries compared with control Vnut^f/f littermates (Fig. 5A). This reduction is not attributed to any changes in total distance traveled during the test (Fig. 5B). Together, these data suggest that female Vnut^GFAPKO mice showed increased anxiety-like behavior. Interestingly, male Vnut^GFAPKO mice did not exhibit any changes in center zone exploration in the same test (Supplemental Fig. 2A, B).

Figure 5. Astrocyte-specific Vnut KO female mice display increased anxiety and depressive-like behavior.

(A) Center zone entries of 4-month-old female Vnut^f/f and Vnut^GfapKO mice during the open field test. Data are shown as mean ± SEM. Two-tailed t-test. *, P < 0.05. N = 12 for Vnut^f/f and 10 for Vnut^GfapKO. (B) Total distance traveled of 4-month-old female Vnut^f/f and Vnut^GfapKO mice during the open field test. Data are shown as mean ± SEM. N = 12 for Vnut^f/f and 10 for Vnut^GfapKO. (C) Percentage of daily consumption of 1% sucrose solution over total liquid consumption in 4-month-old female Vnut^f/f and Vnut^GfapKO mice. Data are shown as mean ± SEM. Two-tailed t-test. *, P < 0.05. N = 12 for Vnut^f/f and 10 for Vnut^GfapKO. (D) Time of immobility of 4-month-old female Vnut^f/f and Vnut^GfapKO mice during the forced swimming test. Data are shown as mean ± SEM. Two-tailed t-test. *, P < 0.05. N = 12 for Vnut^f/f and 10 for Vnut^GfapKO. (E) Time of immobility of 4-month-old female Vnut^f/f and Vnut^GfapKO mice during the forced swimming test with pre-treatment of saline or 20 pmol 2-Me-SATP (i.c.v.). Data of individual mouse with or without 2-Me-SATP treatment are plotted and matched. Two-way RM ANOVA followed by Sidak’s multiple comparisons test. ***, P < 0.001; ****, P < 0.0001. N = 10 for Vnut^f/f and 9 for Vnut^GfapKO.

To assess the depressive-like behavior in Vnut^GFAPKO mice, we performed a sucrose preference test, in which animals were given concurrent choice of water or palatable 1% sucrose solution. As expected, female Vnut^f/f mice consumed more than 80% of sucrose solution daily (Fig. 5C). In contrast, female Vnut^GFAPKO mice showed significantly less sucrose preference (Fig. 5C), demonstrating anhedonia, a key presentation of depressive-like behavior. No other type of malaise was otherwise observed in Vnut^GFAPKO mice.

In a forced swimming test, time of immobility of mice was used to quantify the despair-like phenotype when exposed to an acute inescapable stress. Consistent with previous behavioral alterations, female Vnut^GFAPKO mice showed ~40% increase in time of immobility during the test (Fig. 5D), indicating increased depressive-like behavior. Additionally, male Vnut^GFAPKO mice showed similar degree of increase in time of immobility compared with control Vnut^f/f littermates (Supplemental Fig. 2C). Notably, the increased time of immobility in the Vnut^GFAPKO mice during the forced swimming test was rescued by pre-administration of non-hydrolysable ATP analog 2-Methyl-S-ATP via intracerebral ventricular injections (Fig. 5E). These data strongly indicate that loss of Vnut-dependent ATP release in astrocytes is responsible for the increased depressive-like behavior in Vnut^GFAPKO mice.

Astrocytic Vnut in the nucleus accumbens is critical for suppressing depressive-like behavior in mice.

Given a significant increase of depressive-like behavior in astrocyte-specific VnutKO mice, we sought to explore the potential neural mechanisms contributing to this behavioral alteration. Dopamine signaling in the nucleus accumbens (NAc) is known to mediate reward and has been implicated in depression in humans^43–45. Previous studies have also demonstrated an important role of astrocytes in the NAc^{22, 46}. Therefore, we examined the neuronal activation in the NAc in Vnut^GFAPKO following inescapable stress. Thus, female Vnut^GFAPKO mice and Vnut^f/f mice were subjected to a 6-min forced swimming test or left undisturbed prior to the c-fos screening in the NAc. One hour after the forced swimming episode, Vnut^f/f mice showed a ~3-fold induction in c-fos reactivity in the NAc core and shell (Fig. 6A–C). In agreement with an increased time of immobility, the induction of c-fos⁺ neurons was significantly reduced in both core and shell of the NAc in Vnut^GFAPKO mice (Fig. 6A–C).

Figure 6. Astrocytic Vnut in the nucleus accumbens is a critical mediator for depressive-like behavior in female mice in response to inescapable stress.

(A) c-fos immunostaining of the nucleus accumbens of the 4-month-old female Vnut^f/f and Vnut^GfapKO mice at basal condition and 1 h after forced swimming test. Scale bar: 500 μm. (B-C) Quantification of c-fos+ neurons in NAc core (B) and shell (C) regions. Two-way ANOVA followed by Tukey’s multiple comparisons test. **, P < 0.01; ****, P < 0.0001. N = 4 animals for basal groups and 6 animals for FST groups. (D) Schematics of AAV8-hGFAP-Cre:GFP injection into the nucleus accumbens of adult female mice. (E) Center zone entries during the open field test of 4-month-old female Vnut^f/f mice injected with AAV8-hGFAP-Cre:GFP or GFP alone in the NAc. Data are shown as mean ± SEM. N = 9. (F) Total distance traveled during the open field test of 4-month-old female Vnut^f/f mice injected with AAV8-hGFAP-Cre:GFP or GFP alone in the NAc. Data are shown as mean ± SEM. N = 9. (G) Time of immobility during the forced swimming test of 4-month-old female Vnut^f/f mice injected with AAV8-hGFAP-Cre:GFP or GFP alone in the NAc. Data are shown as mean ± SEM. Two-tailed t-test. **, P < 0.01. N = 9. (H) Percentage of daily consumption of 1% sucrose solution over total liquid consumption in 4-month-old female Vnut^f/f mice injected with AAV8-hGFAP-Cre:GFP or GFP alone in the NAc. Data are shown as mean ± SEM. Two-tailed t-test. *, P < 0.05. N = 9.

To examine the functional significance of astrocytic Vnut in the nucleus accumbens, we performed stereotaxic injections of AAV8 encoding human GFAP promoter-driven Cre:GFP into the nucleus accumbens of adult female Vnut^f/f mice to induce local Vnut deletion (Fig. 6D). AAV8 encoding human GFAP promoter-driven GFP was used as a control. 2 weeks after AAV injection, mice receiving AAV8-hGFAP-Cre:GFP showed similar total distance travelled and center zone entries as control mice receiving AAV8-hGFAP-GFP during the open field test (Fig. 6E, F). During the forced swimming test, mice with the Vnut deletion in the NAc showed a ~2-fold increase in time of immobility than control animals (Fig. 6G). Further, these mice displayed reduced preference for sucrose solution compared with control mice (Fig. 6H), demonstrating anhedonia. Together, these data strongly support that astrocytic Vnut in the nucleus accumbens plays a crucial role in suppressing depressive-like behavior.

Loss of Vnut in astrocytes affects dopamine signaling in the nucleus accumbens.

Over 95% neurons in the nucleus accumbens are dopamine receptor-expressing medium spiny neurons. Reduced neuronal activation may be contributed by a decrease in dopamine release in Vnut^GFAPKO mice. To test this, we used carbon fiber amperometry to measure evoked dopamine release in the NAc of ex vivo cultured brain slices from female Vnut^GFAPKO mice and their Vnut^f/f littermates. The release of dopamine molecules in the NAc was evoked in response to acute electrical field stimulation by a bipolar electrode in acute coronal slices from Vnut^f/f mice (Fig. 7A). On average, the female Vnut^GFAPKO mice showed a 63% reduction (P = 0.0534) in evoked dopamine release in the NAc compared with their Vnut^f/f littermates (Fig. 7A). This was, at least in part, due to a trend toward a reduction in the amplitude as well as the half-life of dopamine molecules in the female Vnut^GFAPKO mice (Fig. 7B, C). This trend was not statistically significant due to a couple of outlier values that could not be excluded from the data analysis.

Figure 7. The effects of astrocytic Vnut loss in adult female mice on evoked dopamine release in the nucleus accumbens.

(A-C) Total numbers (A), amplitude (B), and T_1/2 (C) of dopamine molecules (DA) released after each electric stimulation in the NAc from 4-month-old female Vnut^f/f and Vnut^GfapKO mice. Data are shown as mean ± SEM. Two-tailed t-test. N = 17 for Vnut^f/f and 12 for Vnut^GfapKO. (D) Relative mRNA expression of astrocytes signature genes in the NAc of 4-month-old female Vnut^f/f and Vnut^GfapKO mice. Data are shown as mean ± SEM. N = 6. (E) Relative mRNA expression of genes involved in dopamine synthesis, release, signaling and degradation in the NAc of 4-month-old female Vnut^f/f and Vnut^GfapKO mice. Data are shown as mean ± SEM. Two-tailed t-test. *, P < 0.05. N = 6. (F) Immunostaining and quantification of tyrosine hydroxylase (Th) in the NAc of 4-month-old female Vnut^f/f and Vnut^GfapKO mice. Scale ar: 100 μm. Data are shown as mean ± SEM. N = 4. (G) Immunoblotting showing the protein expression levels of Gfap, Th, MaoA and MaoB in the NAc tissues of 4-month-old female Vnut^f/f and Vnut^GfapKO mice. Gapdh is used as loading control. (H) Densitometric analysis of Gfap and Th in the NAc of 4-month-old female Vnut^f/f and Vnut^GfapKO mice. Data are shown as mean ± SEM. N = 6. (I) Densitometric analysis of MaoA and MaoB in the NAc of 4-month-old female Vnut^f/f and Vnut^GfapKO mice. Data are shown as mean ± SEM. Two-tailed t-test. *, P < 0.05. N = 6.

Despite any likely differences in evoked neurotransmitter release in the nucleus accumbens, loss of Vnut in astrocytes does not impact the expression of any homeostatic and reactive signature genes of astrocytes in the same site (Fig. 7D). The mRNA expression of genes related to dopamine synthesis, transport, and signaling were not altered (Fig. 7E). However, the expression of monoamine oxidase A (Maoa), which is responsible for the turnover of dopamine, was significantly increased in the NAc of the Vnut^GFAPKO mice (Fig. 7E). In addition, Maob mRNA level showed a trend toward increase in Vnut^GFAPKO mice (Fig. 7E). At the protein level, tyrosine hydroxylase (TH) and Gfap were not altered in the NAc of Vnut^GFAPKO mice compared with control littermates (Fig. 7F–H), indicating normal dopaminergic axonal terminal density, dopamine synthesis, and reactivity of astrocytes in the NAc of Vnut^GFAPKO mice. Notably, the protein expression levels of MaoB, but not MaoA, was significantly increased in the NAc of Vnut^GFAPKO mice (Fig. 7G, I), which could contribute to the reduced dopamine signaling in these mice.

Loss of Vnut in astrocytes reduces motivation for reward in mice.

Given a relatively low coverage of GFAP-CreERT2 in astrocytes in adult mice, we next aimed to further confirm the role of astrocytic Vnut in behavioral modulation using an independent astrocytic-specific Cre driver. Thus, we crossed Vnut^f/f mice with Aldh1l1-CreERT2 mice, which has been shown to target astrocytes in adult mice with much higher efficiency⁴⁷. The knockout efficiency of Vnut gene in astrocytes was demonstrated by reverse-transcription PCR on mRNAs isolated from magnetic-activated cell sorted (MACS) astrocytes via astrocyte-specific ACSA-2 antigen (Supplemental Fig. 3). 4-month-old female Vnut^Aldh1l1KO mice exhibit normal center zone exploration and normal locomotor activity during the open field test (Fig. 8A, B). Consistent with the Vnut^GFAPKO mice, the female Vnut^Aldh1l1KO mice displayed significantly decreased sucrose preference (Fig. 5C) and increased time of immobility during the forced swimming test (Fig. 5D). Taken together, our data from two independent astrocytic-specific VnutKO mouse models strongly suggest an important role of astrocytic exocytosis of ATP on suppressing depressive-like behavior.

Figure 8. Loss of Vnut in Aldh1l1⁺ astrocytes increases depressive-like behavior and reduces motivation for reward in adult female mice.

(A) Center zone entries of 4-month-old female Vnut^f/f and Vnut^Aldh1l1KO mice during the open field test. Data are shown as mean ± SEM. N = 9 for Vnut^f/f and 10 for Vnut^Aldh1l1KO. (B) Total distance traveled of 4-month-old female Vnut^f/f and Vnut^Aldh1l1KO mice during the open field test. Data are shown as mean ± SEM. N = 9 for Vnut^f/f and 10 for Vnut^Aldh1l1KO. (C) Percentage of daily consumption of 1% sucrose solution over total liquid consumption in 4-month-old female Vnut^f/f and Vnut^Aldh1l1KO mice. Data are shown as mean ± SEM. Two-tailed t-test. **, P < 0.01. N = 9 for Vnut^f/f and 8 for Vnut^Aldh1l1KO. (D) Time of immobility of 4-month-old female Vnut^f/f and Vnut^Aldh1l1KO mice during the forced swimming test. Data are shown as mean ± SEM. Two-tailed t-test. *, P < 0.05. N = 9 for Vnut^f/f and 10 for Vnut^Aldh1l1KO. (E) Active nose pokes for the active port of 4-month-old female Vnut^f/f and Vnut^Aldh1l1KO mice during the fixed ratio 1 and 5 training sessions for sucrose pellets. Data are shown as mean ± SEM. N = 9 for Vnut^f/f and 12 for Vnut^Aldh1l1KO. (F) Breakpoints of 4 month-old female Vnut^f/f and Vnut^Aldh1l1KO mice during the progressive ratio schedule for each reinforcement. Data are shown as mean ± SEM. Two-tailed t-test. *, P < 0.05. N = 9 for Vnut^f/f and 12 for Vnut^Aldh1l1KO. (G) Representative Nissl staining showing the position of the implanted guide canula in the NAc for microdialysis. (H) Dopamine concentrations in the dialysates collected from 5-month-old female Vnut^f/f and Vnut^Aldh1l1KO mice. Data are shown as mean ± SEM. N = 9 for Vnut^f/f and 11 for Vnut^Aldh1l1KO.

We further examined the motivation for reward in Vnut^Aldh1l1KO mice. Thus, both male and female Vnut^Aldh1l1KO mice and their littermates were trained on a feeding experimentation device (FED3) on fixed ratio 1 (FR1) schedule to obtain palatable sucrose pellets. After 10 days of training, the behavioral performance of all animals stabilized, showing equivalent events of nosepoking for the active portal for sucrose reward (Fig. 8E and Supplemental Fig. 4A). When the task difficulty was raised to fixed ratio 5 (FR5), meaning 5 active nosepokes for one sucrose pellet, both control and Vnut^Aldh1l1KO mice increased the events of active nosepokes during the session (Fig. 8E and Supplemental Fig. 4A). However, female Vnut^Aldh1l1KO mice showed a slightly, but not significantly reduced nosepoking activity for the active portal (Fig. 8E). Importantly, when switched to progressive ratio (PR), during which exponentially more nosepokes on the active portal was required for the next reinforcement, female Vnut^Aldh1l1KO mice gave up much sooner, showing a significant decrease in PR breakpoints (Fig. 8F). In contrast, male Vnut^Aldh1l1KO mice displayed similar breakpoint as Vnut^f/f males (Supplemental Fig. 4B). These behavioral alterations strongly support a reduction of motivation for reward in female Vnut^Aldh1l1KO mice. Given the strong correlation between dopamine signaling and motivation^{48, 49}, both Vnut^Aldh1l1KO and Vnut^f/f littermates were subjected to in vivo brain microdialysis in freely moving state for extracellular dopamine assessment. Consistent with the observed behavioral changes, the basal extracellular levels of dopamine in the nucleus accumbens in the female Vnut^Aldh1l1KO mice were reduced by ~59% compared with control Vnut^f/f littermates (Fig. 8G and H). Together, these data suggest that loss of Vnut in astrocytes leads to increased depressive-like behavior and reduced motivation for reward, due to a reduction of dopamine signaling in the nucleus accumbens in female mice.

Effect of astrocyte-specific Vnut knockout on cognitive and social behavior.

Since the Aldh1l1-CreERT2 driver efficiently targets astrocytes in the cortical and hippocampal regions⁴⁷, we also assessed cognitive and social behavior in Vnut^Aldh1l1KO mice. To examine working memory, both 5-month-old male and female Vnut^Aldh1l1KO mice and their Vnut^f/f littermates were placed in a Y-maze for free exploration. The total distance traveled and the number of arm entries between Vnut^f/f and Vnut^Aldh1l1KO mice were comparable (Supplemental Fig. 5A, B, D, E). Importantly, the spontaneous alternation of arm entries between Vnut^f/f and Vnut^Aldh1l1KO mice was not different (Supplemental Fig. 5C, F), indicating normal working memory of mice following Vnut deletion in astrocytes.

We also assessed the short-term object recognition/memory of astrocyte-specific VnutKO mice. To this end, animals were allowed free exploration of two identical objects, followed by a second session 2 min later with one object replaced by a novel object. Total time of interactions with each object was quantified and compared between the novel and familiar object. Vnut^f/f and Vnut^Aldh1l1KO mice exhibited the same degree of preference on the novel object in both males and females (Supplemental Fig. 5G, H). Therefore, loss of Vnut in astrocytes does not impact object recognition and short-term memory in mice.

In a 3-chamber social preference test, when given the choice of a novel object or an unfamiliar mouse of the same sex, both male and female Vnut^f/f mice preferred to interact with the unfamiliar mouse, indicating social preference as expected in normal mice (Supplemental Fig. 5I, J). Interestingly, the male Vnut^Aldh1l1KO littermates showed significantly more preference for social interactions compared with male Vnut^f/f mice (Supplemental Fig. 5I), whereas female Vnut^Aldh1l1KO mice displayed similar degree of social preference as the female Vnut^f/f littermates (Supplemental Fig. 5J). Taken together, loss of Vnut in astrocytes increases depressive-like behavior and impairs motivation in female mice without a major impact on cognition and social behavior.

Discussion

Astrocytes have emerged as a key player in the development and plasticity of neural circuitry. The homeostatic functions of astrocytes include regulation of neurotransmitter re-uptake, extracellular pH and ion concentrations, synaptic plasticity, blood-brain barrier integrity, and brain metabolism². At least some of these important functions of astrocytes are mediated by the release of signaling molecules. Thus, astrocytes have been shown to release neurotransmitters like glutamate, D-serine and ATP, each of which play important roles in modulation of synaptic plasticity¹⁶.

In the present study, we investigated the functional significance of astrocyte exocytosis of ATP on motivational behaviors and dopaminergic neurotransmission. We specifically targeted the vesicular nucleotide transporter (Vnut) in astrocytes, since Vnut is responsible for loading cytosolic ATP into the secretory vesicles in cells, and loss of Vnut in astrocytes would effectively eliminate astrocytic ATP release via exocytosis. We show that loss of Vnut significantly decreases the loading and release of ATP or its florescent analog Mant-ATP in primary astrocytes. Loss of Vnut in astrocytes has little effect on the expression of most homeostatic genes in astrocytes or on the reactivity of astrocytes, as represented by Gfap expression. Using genetically engineered mice carrying floxed Vnut allele and the astrocyte specific inducible GFAP-CreERT2 mice and Aldh1l1-CreERT2 mice, we show that knockout of Vnut in astrocytes, especially in females, leads to significantly increased depressive-like behavior and reduced motivation for reward, suggesting decreased dopamine signaling in the nucleus accumbens (NAc), the reward center in the brain. Evoked dopamine release in the NAc of coronal brain slices from astrocyte-specific Vnut^GFAP knockout female mice was reduced, although this did not quite reach significance due to a couple of outlier mice. Importantly, however, an independent cohort of astrocyte-specific Vnut^Aldh1l1 knockout female mice displayed significantly reduced basal extracellular dopamine levels in the nucleus accumbens. Furthermore, there was a significant increase in the expression of monoamine oxidases. Together, our study indicates that astrocytes release ATP via exocytosis to modulate motivation for reward through a likely impact on dopaminergic axonal terminals in the NAc.

Extracellular ATP may be derived from different molecular routes, including cell death, activation of exocytosis, and opening of non-specific channels, such as connexin-composed hemichannels^50–52. Astrocytes are equipped with key molecular components for ATP release via all these routes. In this study, using genetic targeting of Vnut in astrocytes in vitro and in vivo, we show that exocytosis contributes a significant portion of ATP release in astrocytes. Importantly, the loss of astrocytic ATP release through this mechanism has functional impact on the dopaminergic neural circuit in adult mice. Thus, astrocytic Vnut KO mice show reduced extracellular dopamine levels in the NAc and reduced motivation for reward, indicating an important effect of astrocyte-derived ATP on dopamine release. This is in agreement with previous studies by our group and others that have shown that ATP analogs can greatly potentiate evoked dopamine release^{22, 53}. Together, these studies indicate that exocytosis of ATP from astrocytes may be one of the key mechanisms contributing to the anti-depressive effect of ATP.

While our data support the functional importance of exocytosis as the source of astrocyte-derived purinergic signaling, the downstream actions of released ATP molecules is likely complex. The half-life of extracellular ATP is very short. After release, ATP is quickly hydrolyzed to ADP and AMP by ecto-ATPases⁵⁴, which can be further converted to adenosine by ecto-5’-nucleotidase⁵⁵. Both ADP and adenosine can act on different types of purinergic and adenosine receptors with different binding affinities and elicit distinct intracellular signaling cascades in diverse target cells. In our study, direct administration of non-hydrolysable ATP analog 2-Me-SATP can rescue the depressive-like behavior in female mice, indicating that ATP acts directly on P2 receptors to elicit the anti-depressive effect. This agrees with other work showing the potentiation effect of other non-hydrolysable ATP analogs on evoked dopamine release²². Also, in agreement with these data, genetic or pharmacologic inhibition of the ecto-ATPase Entpd1 (CD39) increases extracellular ATP levels in the hippocampus and elicits anti-depressive effect in mice subjected to chronic social defeat stress⁵⁶.

In addition to ATP, adenosine is another key component of purinergic neurotransmission and important for the modulation of various neural circuits. Thus, astrocyte-derived adenosine plays a critical role modulating sleep via adenosine A1 receptor⁵⁷. Adenosine signaling also impacts emotional behavior. Inhibition of ecto-5’-nucleotidase (CD73) reduces the conversion of AMP to adenosine and attenuates synaptic dysfunction and mood dysfunction^{58, 59}. Furthermore, adenosine has been shown to tonically suppress dopamine release⁶⁰. Collectively, these studies suggest that increasing extracellular ATP, while suppressing adenosine production, could support dopamine neurotransmission and suppress emotional deficits.

Astrocytes are enriched for a unique set of SNARE complex proteins required for regulated exocytosis, including the plasma membrane bound syntaxin-4 and SNAP-23, and the vesicular membrane bound Vamp3 and Vamp7 particularly on secretory lysosomes^61–64. Importantly, Vnut is highly enriched in the secretory lysosomes in astrocytes⁵⁰. The expression pattern of these key components indicates that exocytosis of ATP in astrocytes is subjected to specific regulations that differ from other neuronal and glial cell types in the brain. Exocytosis in astrocytes is tightly coupled with surrounding synaptic activity, as extracellular neurotransmitters like glutamate and ATP can induce intracellular calcium spikes in astrocytes to activate SNARE protein-mediated exocytosis. In addition, exocytosis of ATP in astrocytes can be stimulated by insulin²². This is interesting because insulin is not known to induce robust intracellular calcium signaling. Instead, insulin induces the tyrosine phosphorylation of two key tyrosine residues on the syntaxin binding protein Munc18c in astrocytes. This phosphorylation alleviates the inhibitory effect of Munc18c on syntaxin-4 by exposing the docking site on syntaxin-4 for Vamp protein recruitment²². Collectively, two tandem signaling events appear to be required for a maximal activation of ATP exocytosis in astrocytes: 1) Munc18c tyrosine phosphorylation to allow Vamp3/7-containing vesicles to dock on syntaxin-4; and 2) elevation of [Ca²⁺]_i to trigger the final membrane fusion and vesicular content release. While future studies are needed, it is plausible that other trophic factors or hormones may also elicit Munc18c tyrosine phosphorylation and regulate exocytotic machinery in astrocytes. Together, with a unique set of SNARE proteins that differ from other neuronal and glial cells in the brain, astrocytes are able to release signaling molecules corresponding to a variety of chemical signals and trophic factors. Notably, hormones like insulin have also been shown to elicit robust transcriptional regulation of various metabolic pathways in astrocytes, including regulation of genes involved in autophagy, cholesterol biosynthesis, and fatty acid biosynthesis⁶⁵. Together these signaling and metabolic changes in astrocytes coordinate the homeostatic regulations in the brain. Further investigation as to how these key molecular pathways involved may help us better understand the comorbidity between neurological diseases and metabolic diseases^{66, 67}.

Data from the present study demonstrate a crucial role of astrocytic exocytosis-derived purinergic signaling in the modulation of dopaminergic neural circuit and behavior. In particular, this astrocyte-dependent modulation occurs at the nucleus accumbens, the projection site for dopaminergic axons. This modulation may involve both direct and indirect mechanisms. Thus, ATP released from astrocytes can directly act on the presynaptic P2Y receptors to modulate dopamine release in the nucleus accumbens^{22, 53}. Alternatively, ATP may act on cholinergic interneurons in the NAc, which have been shown to regulate dopamine release and decision making^{68, 69}. Purinergic signaling may also affect the dopamine signal turnover after synaptic release. Supporting this concept, our data show that the protein expression of monoamine oxidases, especially Maob, is significantly increased in astrocyte-specific VnutKO mouse brains. Similar molecular alterations, supplemented by reduced evoked dopamine release, have also been observed in brain-specific insulin receptor KO mice, which display similar depressive-like behavior⁷⁰. While the present study demonstrates that loss of Vnut in astrocytes in NAc is sufficient to induce depressive-like behavior in mice, the potential functional impact of astrocytic Vnut in other brain regions related to emotion and reward should not be overlooked. In addition to modulating distal axonal terminals in the NAc, astrocytes are known to modulate the dopaminergic neural circuits originating in the ventral tegmental area (VTA), where the dopamine cell bodies reside. Selective activation of astrocytes in the VTA by chemogenetic approaches is sufficient to increase the burst firing of dopamine neurons⁷¹. Whether exocytosis of ATP in VTA astrocytes also play important roles in modulating the excitability of dopamine neurons awaits further investigation.

It is now clear that astrocytes are capable of releasing various neurotransmitters through exocytosis to modulate neural circuits. The types, organizations, distributions, and release patterns of vesicles containing different types of neurotransmitters in astrocytes, however, are complex. For instance, inorganic polyphosphate (PolyP) is present in Vnut⁺ vesicles in astrocytes, but not in vGlut-containing vesicles, which are primarily equipped for glutamate release⁷². Given PolyP’s ability to activate purinergic receptors⁷³, the co-release of PolyP and ATP from Vnut-expressing vesicles may have a combined synergistic effect to activate purinergic receptors. While our present study focuses on Vnut and purinergic signaling, astrocytes also express necessary components for regulated vesicular exocytosis of glutamate, including the vesicular glutamate transporters (vGlut1 and vGlut2) that are specifically responsible for vesicular loading of glutamate⁷⁴. Astrocyte-derived glutamate signaling plays critical roles in modulating various neural circuits in the brain^{75, 76}. Whether Vnut-dependent ATP signaling and vGlut-dependent glutamate signaling interact with each other is not yet understood. Studies have shown that Vnut and vGlut co-exist in a subset of synaptic vesicles in neurons⁷⁷, suggesting a co-release of ATP and glutamate and a possible interaction between the two signaling cascades in neurons. Further investigation awaits to confirm whether astrocytes contain Vnut⁺/vGlut⁺ vesicles similarly to neurons. It is possible that in astrocytes, loss of Vnut could affect vesicular loading and release of glutamate, and eventually affect neural circuit modulation, including dopamine release in the mesolimbic system we report in the present study.

In summary, Vnut is responsible for a significant portion of ATP release from astrocytes, loss of which leads to functional consequences including increased depressive-like behavior, and decreased motivation for reward. Our study demonstrates a potentially important role of astrocyte-derived purinergic signaling in dopaminergic and, more inclusively, monoaminergic neurotransmission. Notably, the modulatory functions of astrocyte-derived purinergic signaling are not limited to dopamine systems and may impact dopamine systems also through changes in glutamate clearance as our results with glutamate transporters suggest. Furthermore, ATP from astrocytes plays a critical role in regulating the sleep cycle, implicating a role for serotonin and its biosynthetic product melatonin. Excess ATP release from astrocytes has also been shown to induce pain hypersensitivity. Together, these studies highlight astrocytes as the powerful “neuromodulatory” glia by acting on multiple nodes within the neural circuitry via distinct molecular mechanisms. Further understanding of the regulations and consequences of astrocyte-derived purinergic signaling in different pathological conditions may provide valuable insights into potentially novel therapeutic development in neurological and psychiatric diseases.